Skip to content
GET IN TOUCH

Overview

Modern vehicles are evolving into complex computing platforms, with increasing software demands and a growing number of Electronic Control Units (ECUs). To reduce complexity, weight, and cost, automotive architectures are shifting toward domain-based consolidation, supported by gateways and high-performance central units.

However, traditional monolithic software is no longer up to the task—it lacks the flexibility, modularity, and safety isolation required for next-generation systems. That’s where containerization and virtualization come into play. By adopting microservices and virtualization technologies, we can create agile, secure, and scalable platforms that enable the Software-Defined Vehicle (SDV) paradigm.

 

The VORTEX-CoLab Approach

To explore this transition, VORTEX-CoLab has developed a prototype that integrates static partitioning hypervisors, Kubernetes-based orchestration (via K3s), and containerized microservices. Our goal: to demonstrate how real-time and non-real-time applications can safely coexist on embedded automotive platforms.

In a recent feasibility study, we validated this architecture for one of our associates. We showed how containers can be deployed in virtual machines (VMs) based on their criticality—even without a fully mature mixed-criticality orchestrator (which we are actively developing as part of one of our publicly funded research projects).

shape

Challenge

Vehicle platforms must support a mix of real-time, safety-critical tasks and high-performance, general-purpose applications—on the same hardware. This creates challenges around resource isolation, scheduling, and system updates. Cloud-native tools aren’t enough: embedded systems demand hardware-assisted virtualization and predictable execution, especially when safety is on the line.

Balancing container orchestration with the strict guarantees required in embedded environments takes careful design. It also requires bridging the gap between modern DevOps tooling and safety-certified platforms.

Solution

VORTEX-CoLab developed a reference architecture and prototype to validate how microservices and real-time applications can coexist safely on embedded automotive platforms. The solution combines static-partitioning virtualization, containerized workloads, and lightweight orchestration techniques adapted for mixed-criticality environments.

Key elements of the solution include:

  • A static-partitioning hypervisor (based on Bao), which provides strong isolation and deterministic execution for real-time and safety-critical workloads.
  • Use of virtual machines (VMs) to separate workloads by criticality level, allowing for a clear separation between hard real-time, soft real-time, and general-purpose applications.
  • Allows containers to run inside VMs, enabling modular and portable deployment of services such as perception, AI inference, and infotainment.
  • Manual orchestration with tailored scheduling configurations, using mechanisms like node affinity, taints, and tolerations to direct workloads to the appropriate VMs based on their resource and isolation requirements.
  • Hardware-assisted virtualization, ensuring that critical applications maintain real-time guarantees while minimizing latency and overhead.
  • A domain-based architecture, which consolidates multiple ECU functions into fewer, more capable embedded compute units, reducing system complexity without compromising safety or scalability.

Although native support for mixed-criticality scheduling in orchestration frameworks is still under development, this architecture demonstrates a practical way to achieve workload separation and safety using current tools and custom configurations. It establishes a solid foundation for future enhancements, including a dedicated mixed-criticality container runtime as envisioned by the SOAFEE initiative.

Use cases

  • Automotive – Software-Defined Vehicles (SDVs): Supports real-time driving logic, over-the-air updates, and infotainment services on a unified platform.
  • Embedded Edge Computing: Enables AI-driven features and real-time control in industrial or robotic environments.
  • Avionics & Industry 4.0: Offers modular system design, isolation, and integration of heterogeneous workloads.
  • High-Performance Embedded Systems: Combines safety, scalability, and serviceability with reduced hardware footprint.

Results

  • Successfully demonstrated coexistence of soft real-time, hard real-time, and general-purpose applications on the same hardware.
  • Built a prototype using the lean, static partitioning hypervisor Bao and container orchestration via K3s.
  • Deployed real automotive based microservices across VMs to reflect real-world mixed workloads.
  • Validated safe and predictable performance through benchmarks and integration with the CARLA simulator.
  • Established a foundation for the future development of a mixed-criticality container runtime.
Need more information about this project?
Contact us
Privacy Overview

This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.